The domain that burned on day twelve

On a Tuesday I'd rather not repeat, two domains went dark inside forty minutes. Both were warming. Both had live sequences queued. I found out from a bounce report, not a platform alert.

Here's what happened, what I did, and what I'd change.

The setup

Two domains, six inboxes, warming toward a first real send on day fourteen. Standard pattern: start at one warmup email a day per inbox, scale by one each morning, watch reputation like a hawk. We were on day twelve. Reputation was fine, opens were normal, no complaints logged, everything on track for a clean launch.

The failure

The registrar deactivated both domains the same morning. I won't get into why. The outcome: MX records evaporated, SPF and DKIM records disappeared with them, every warming inbox went offline, and every scheduled send queued then bounced. The warmup timer froze at day twelve.

The first thing I noticed was not the domains themselves. It was a bounce report from Instantly showing an unusually high failure rate on morning sends. I thought it was a provider issue and opened the registrar dashboard to double-check the records. The domains were gone.

The wrong thing to do

Panic, buy replacement domains, point them at the same inboxes, and start sending on day one. I have watched other operators do exactly this, in real time, in a group chat. It is a faster way to get blacklisted than letting the original damage play out.

The logic people use to justify it goes like this: "The inboxes are still warm, we just need domains pointed at them." That is not how receiving mailservers think about reputation. Reputation lives with the domain, not with the inbox. A new domain sending from day one, even with a "warm" inbox underneath, is a new domain sending from day one.

What we actually did

I bought three fresh domains inside four hours, none of them variants of the original name. Spaced the purchases across three different registrars, because the whole point of what just happened was single-registrar risk.

We provisioned six inboxes across the three domains, not concentrated on one, then set the warmup to start on day three, not day one. Gave the registrars time to propagate DNS before we did anything.

We sent no prospect email for nine days.

The lesson

Single-registrar risk is real and I was taking it. That is the first thing.

The second: every agency running cold email needs a domain bench. At least two unused, fully warmed domains sitting idle, ready to slot in. Warmed, not just bought. A domain purchased the day you need it is useless to you. We now run four production domains and two bench domains. When a production domain shows stress, we rotate to the bench and start warming a replacement. Continuous rotation.

A question we ask every prospect now

"If your current outbound domain died tomorrow, how many days of pipeline would you lose?"

Most founders cannot answer. That is the correct reason to be concerned, because if you can't answer, the answer is a lot.